Paper Round-up: Pacific Conference for Development Economics (PacDev) 2020

These short summaries of talks delivered at PacDev 2020 were written by CEGA staff with an introduction by Bilal Siddiqi, CEGA’s Director of Research.

On Saturday, March 14th, 2020 CEGA hosted the Pacific Conference for Development Economics (PacDev), the largest conference on Development Economics on the West Coast, and one of the leading annual Development Economics events in the United States. The original call for papers garnered 270 submissions, from which we selected 70 speakers, who presented on a diverse array of research topics. We were also grateful to have UC Berkeley Economist Gabriel Zucman keynote, with a talk on “21st Century Tax Systems for Developing Countries.”

Amidst the rapidly evolving situation with COVID-19, we decided to transition to a fully online conference just a week out from the event. Thankfully, the transition allowed for greater participation, attracting 340 attendees over the course of the day. This was almost twice as many as had been expected to attend the in-person conference! For more about this adaptation and take-aways from hosting a virtual conference, see our companion blog post (forthcoming).

We hope the short summaries of each talk, organized by session below, are useful. Enjoy!

CEGA staff facilitate PacDev 2020 from our “Control Room” on campus, just days before strict social distancing measures were put in place.

Session 1A: Digital Technology

• Access to cell phone coverage leads to a 10.8 percentage point reduction in the likelihood that a village has an Ebola Virus Disease case during the 2014 outbreak in Liberia. (Gonzalez and Maffioli)
• Social connectivity and household expenditures increased when community cell towers were built in isolated villages in the Philippines, but there was no impact on remittances, informedness, or market access. (Blumenstock et al) *working paper forthcoming
• An app connecting agricultural laborers to employers both motivated employers paying above-average wages to pay less, and employers paying below-average wages to pay more, compressing wages toward the average, and suggesting some reduction in the cost of searching for a job. (Jeong)

Session 1B: Gender Discrimination

• In a lab-in-the-field experiment in Ethiopia, subjects are ten percent less likely to follow the same advice from a female leader than an otherwise identical male leader, and female-led subjects perform .34 standard deviations worse as a result. (Ayalew et al)
• When female sellers in India are provided with the same market stall and goods as male sellers, female sellers perform the same as their male peers in profitability. This points to access to capital, rather than buyer-side discrimination, as the cause of lower revenue for female-owned businesses. (Delecourt and Ng)
• What is the impact of migration restrictions on women’s fertility in Sri Lanka? Young women, who are restricted from migrating, increase their fertility; while older women, who are restricted from migrating only if they have young children, reduce their fertility. As a result, new mothers are less-educated and younger, with potential impacts on children. (Peru)

Session 1C: Credit Markets

• An analysis of the effect of cash shortages following India’s demonetization policy found that regions with fewer banks had lower economic activity and voters reported having less favorable views on demonetization, leading the ruling party to do relatively worse in said regions’ next election cycles. (Khanna and Mukherjee)
• Research on repayment flexibility in microfinance contracts in Uttar Pradesh, India shows that borrowers who are allowed to choose between a standard, rigid contract or a more expensive, flexible contract experience better business outcomes (but similar repayment rates) than if they are just given a rigid contract. (Barboni and Agarwal)
• An evaluation of a new credit product launched in Paraguay suggests that granting loans using new technologies to construct alternative screening mechanisms may be particularly beneficial when targeted to clients with limited credit market experience. (Azevedo et al)
• Research from Uganda shows mixed effects when a formal financial institution lends to savings groups and lets the group decide the allocation of borrowed funds. Borrowers had lower rates of food insecurity in the short-run, but the selection effects from the lending led to more turnover among groups assigned to loans. (Burlando et al)

Session 1D: Education

• Parental out-migration in rural China negatively affects math and language scores of children who are left behind. (Luo)
• A lab-in-the-field experiment in Malawi measures parents’ preferences for allocating resources across their children finding that while parents place some weight on maximizing earnings, they also display a strong preference for equality in inputs. (Berry et al)
• Looking at middle school kids in Sao Paulo, this paper finds that there are causal peer effects for kids’ aspirations to go to college — an extra friend aspiring to go to college increases the likelihood that a student will also aspire to it by 9.9% on average, and decreases the likelihood of dropping out of high school. (Miranda)
• This paper examines the effects of a new skills-based curriculum implemented by the Early Years Preschool Program in Bangladesh. The authors find large intent-to-treat effects on children’s cognitive and socio-emotional development, along with significant impacts on parental investment measures. (Rodriguez and Saltiel)

Session 1E: Behavior & Motivation

• This paper finds that students with a lower socio-economic status perform worse on mathematics questions that are about money, relative to their peers. (Duquennois)
• Studying HIV patients at a clinic in Uganda, this study finds that commitment to a goal and subsequent effort to achieve it is increased when the participant perceives it as attainable. (Huang and Linnemayr)
• While appointments for preventative care are common in developed countries, they are not as common in countries like Malawi. This paper finds that appointment-setting more than doubles HIV testing rates for high-risk men. (Derksen et al)
• In Ecuador, undernutrition for children under five is as high as 24 percent. This study looks at the effect of text messages to caregivers that encourage regular health checkups, timely integration of solid foods, diet diversity, and good water and sanitation practices, and finds that messages lead to a 23 percent decline in symptoms of common illnesses. (Rounseville et al)

Session 2A: Mobility & Long-Term Development

• Given deworming’s low cost, deworming provides an extremely high social internal rate of return. Looking at the 20-year impacts of deworming in rural Kenya, this paper finds that the treatment group experiences a 10% gain in consumption expenditure, a 7% gain in total adult earnings, and is more likely to live in urban areas. (Baird et al)
• This paper uses a large dataset of 3000 cities and over 100,000 urban neighborhoods to explore residential segregation in urban India of Scheduled Castes/Scheduled Tribes (SC/STs) and Muslims. Compared to U.S. racial black/non-black segregation, segregated cities in India have worse access to schools, doctors and public hospitals. (Adukia et al)
• Subsidized housing lotteries in Mumbai allocate housing to low and middle-income residents. This paper identifies the human capital effects of such a transfer, showing that 3–5 years later, winners are more educated than non-winners, with effects concentrated among school-age youth. (Kumar)

Session 2B: Taxes & Public Goods

• Constituencies with “dishonest politicians” in Bangladesh (i.e. those who have evaded income taxes), are more likely to have lower public goods provision, including school enrollment and social safety net measures, than constituencies with honest politicians.(Mahzab)
• Using a field experiment in the DRC, this paper estimates the elasticity of tax compliance and tax revenue with respect to the tax rate, finding that the government could increase revenues by lowering tax rates. (Bergeron et al)
• In a large city in Haiti, this ongoing project cross-randomizes trash collection, property tax invoice delivery, and public exposure of tax compliance (exterior stickers posted on complying buildings) over 50,000 properties, and follows responses in the government’s administrative data. (Krause)

Session 2C: Political Economy

• This study finds that indigenization (local takeover) of oilfields in Nigeria increases output, mainly by mitigating conflict risk, theft of oil, and violent attacks by criminal-militant groups through local connections to politicians and the army. (Rexer)
• Overall, faster courts result in higher GDP per capita. This study finds that in India, reducing judge vacancy by adding one judge reduces the backlog of cases by 6%, increases bank lending by .5%, and increases firm profits by 1.6% leading to an overall 8:1 benefit to cost ratio. (Rao)
• Do municipalities run by a member of the presidential party receive more government credit? This paper finds that in Mexico, these municipalities are 7.9% more likely to obtain public credit and receive 58.5% more credit per capita while finding no differences in the allocation of private credit. (Garza and Lopez-Videla)
• This paper explores the relationship between crime and housing values, in particular, the impact of new police stations on crime rates and housing costs. Researchers find that an observable police presence reduces violent crime by 20% and property crime by 40%, and increases housing values in areas surrounding new police stations by 6%. (Morales-Mosquera)

Session 2D: Trade & Market Integration

• A study using India’s demonetization episode as a natural experiment finds that financial infrastructure is an important determinant of whether interventions can force households to switch from informal to formal financial transactions over the long run. (Aggarwal et al)
• Researchers find that India’s efforts to increase foreign capital liberalization in the 2000s reduced capital misallocation and increased productivity. (Bau and Matray)
• Evidence from Ethiopia suggests that rural road construction increased market integration and household consumption, primarily through new trading opportunities. (Kebede)
• Data gathered from the introduction of border facilities in Uganda shows that the incidence of formal border-crossings for trade is affected by taxes, gender, and time-sensitivity, which reflects the persistence of informality in border crossing choices. (Siu)

Lightning Talks:

• A randomized evaluation on the effect of gender integration in the workplace in India found zero effects on productivity and days present during the study period, but an overall increase in knowledge sharing, dating, and comfort with the opposite gender for male employees. (Batheja)
• An evaluation of the effect of women’s job opportunities on their ability to choose their spouse found that an increase in women’s labor demand at the time of marriage increases their ability to independently choose their spouse, with additional findings related to religion, education-level, and location. (Chatterjee and Heath)
• Examining the effects of macroeconomic shocks on child schooling in Turkey, researchers found that higher unemployment among unskilled workers increases schooling, whereas higher unemployment among skilled workers decreases schooling. (Gunes and Marchand)
• By examining marriage market outcomes of displaced women during the Partition of India in 1947, researchers were able to determine that women who were young were significantly more likely to marry as children relative to displaced women who were older. (Bharadwaj et al) *working paper forthcoming
• Researchers gathered data across four countries in Africa to study the size, source, and persistence of productivity dispersion among smallholder farmers, with results that question some common implications of observed dispersion. (Maue et al)
• Using Peru’s civil war in the 1980’s to measure the impact of exposure to conflict at different stages of childhood, this evaluation found evidence that exposure to conflict among primary school aged increased their probability of incarceration in adulthood. (Sara)

Keynote: Gabriel Zucman

Gabriel Zucman of UC Berkeley keynoted this year’s PacDev, with a talk on “21st Century Tax Systems for Developing Countries” (click below to view the video). Zucman is an Assistant Professor of Economics, where his research focuses on the accumulation and distribution of global wealth.

Session 3A: Measuring Poverty & Welfare

• Combining mobile phone transaction data with household survey data from rural Afghanistan, researchers study the extent to which machine learning methods can accurately estimate different measures of poverty and deprivation. Findings show that machine learning methods have little predictive power for estimating an asset-based wealth index or expenditures, but can (to some degree) differentiate ultra-poor from non-ultra-poor households. (Aiken and Blumenstock) *working paper forthcoming
• Researchers from IDinsight estimate the value of statistical life (VSL) in low-income populations of Kenya and Ghana through a stated preferences survey, finding that respondents placed a uniquely higher value on averting deaths and averting the death of individuals under 5. (Redfern et al)
• This paper uses machine learning and causal inference to predict the marginal effects of a program on a recipient, making it possible to infer the policy’s implied preferences over households and outcomes. This methodology is then used to show the implicit priorities of Mexico’s PROGRESA program. (Björkegren et al)
• Using machine learning and causal inference to predict the impact of the Green Revolution in India on nutrition and cropping patterns, researchers find a total increase in women’s height and health and increased acreage to rice and wheat (over coarse cereals and pulses). (Bevis and Negi)

Session 3B: Energy & Environment

• An agricultural extension program in Uganda offering inputs and training reduced deforestation as farmers increased intensification of land use. (Abman et al)
• Random variation of a voucher program to cover fees for household electrification in El Salvador increased adoption significantly, and electrification led to a 56% increase in female participation in non-farm employment and an extra $450 dollars per year in income. However, there was no impact on the male labor supply. (Barron and Torero)
• Electricity meter installation in the Kyrgyz Republic showed no changes to electricity theft, but improved consumer welfare by substantially impacting the quality of service through reductions in outages and voltage spikes, and increases in repairs. (Meeks et al)

Session 3C: Gender-Based Violence

• Using a novel dataset of 33 ethnic civil conflicts in Africa and their use of sexual violence, this paper finds that gender-unequal armed actors are more likely to be perpetrators of sexual violence, and and that sexual violence increases when the perpetrator is more gender-unequal than the victim. (Guarnieri and Tur-Prats)
• In Jordan, women’s attitudes toward intimate partner violence (IPV) are affected by the amount that the groom paid to the bride upon marriage, and the bride’s contribution to the marriage costs: both payments are associated with a positive and significant increase of women’s likelihood to justify IPV. (Khalifa)
• This paper investigates the impact of protests after a gang-rape incident in Delhi (India) on a moving bus in December 2012 on reporting of violence against women (VAW). Regions that were more exposed witnessed a sharp increase in reported VAW, and the increase in reported VAW can be attributed to a rise in reporting rather than actual incidence. (Sahay) *working paper forthcoming

Session 3D: Local Institutions

• This paper studies the effects of the sub-Saharan African “Matrilineal belt” — where kinship and inheritance is traced through women — on well-being finding that children of matrilineal women are healthier, better educated, and experience less domestic violence, resulting mainly from improvements to girls’ education. (Lowes)
• Studying dairy cooperatives in Karnataka, India, this paper finds that incentives for improved production quality are effective, but the structure of the group payment affects the quality more than the size of incentive. (Shenoy and Rao)
• This paper finds that deforestation in India was reduced by a law passed in 1996 which elected local government councils to monitor deforestation rates. These effects increased in areas with greater mining potential, and where ethnic groups with strong resource-preservation norms are the majority. (Lal et al)
• Drawing from a lab-in-the-field experiment in the DRC, this paper finds that Christian missions (a) result in less bias against out-group members and (b) are associated with the adoption of less communal and more universal moral values. (Bergeron)

Session 4A: Jobs

• A randomized experiment benchmarking a youth employment training program against cash grants in three Rwandan districts aims to provide a direct cost-equivalent comparison of in-kind development aid to cash. (McIntosh et al) *working paper forthcoming
• The evaluation of a government-sponsored apprenticeship training program in Ghana found that there was usually a short term loss when shifting from wage work to self-employment post-apprenticeship. (Hardy et al)
• Job displacement in Medellin, Colombia led to higher probabilities of arrest for both the displaced worker and youth in the family; areas with more opportunity for reemployment saw smaller increases in the probability of arrest. (Khanna et al)
• Trade-induced job displacement in Mexico had heterogeneous impacts on mortality: increased mortality from diabetes, linked to raised obesity rates and less access to nutritious food, was offset by declines in mortality from alcohol-related liver disease and ischemic heart disease. (Fernandez Guerrico)

Session 4B: Climate Change & Pollution

• Programa de Agua Limpia (PAL) in Mexico has a persistent impact on cognitive and physical development (~6% increase in cognitive assessment score and .11 standard deviation increase in height in adolescence), with effects on human capital that persist to at least to early adulthood and lead to increased hourly earnings. (Bhalotra et al)
• Climate change is leading to higher rates of malnourished children in sub-saharan Africa. There’s a negative relationship between weight and temperature at temperatures above 25 degrees celsius due to lower agricultural yields in rural areas, and higher rates of diarrheal diseases in urban areas. (Baker and Anttila-Hughes)
• Research already suggests climate change is leading to more violent conflict, but using an opportunity cost model for agricultural income, it appears that while conflict occurs during bad harvest years, conflict may decrease if people assume that future droughts will be more intense. (Roche et al)

Session 4C: Provider Incentives

• This paper explores what encourages or discourages reporting peers’ absenteeism amongst employees of the Afghan Ministry of Education. Adding a monetary incentive to report absent colleagues actually reduces the rate of reporting, as individuals are morally averse to being paid for harming their peers. (Fiorin)
• Teachers have significantly more information about their quality than their principal and when given the option to be paid by a performance pay structure or a fixed rate structure, only high value add teachers select into the performance pay structure suggesting that the benefits of performance pay contracts on teaching quality are underestimated. (Andrabi and Brown)
• This paper explores the causal effect of organizational mission on performance of community health workers in Pakistan. Workers visit more households and improve the quality of service after participating in mission development workshops. This improvement is roughly half of what a performance-linked financial incentive can achieve. (Khan) *working paper forthcoming
• This paper provides the first large scale evidence of private hospital responses to changes in reimbursement rates under public health insurance in India finding (a) substantial coding manipulation in response to increased reimbursements, and (b) pervasive informal out-of-pocket charges although services are supposed to be free. (Jain) *working paper forthcoming

Session 4D: Risk

• Using two-panel data sets from Indonesia and Mexico (n=15,000), the authors find that living through periods of increasing macroeconomic volatility increases measured risk aversion and decreases risk-taking behaviors like smoking, migration, and the planting of cash crops. The results hold even when controlling for changes in subjects’ economic circumstances, suggesting psychological adaptation to the experienced volatility. (Levin and Vidart)
• This paper shows that the consumption pattern in rural Tanzania is consistent with a self-insurance regime, and that risk aversion varies substantially across districts. Imposing a strict condition on interest rates, as often done in prior literature, misses their intertemporal heterogeneity and biases the estimation of risk aversion. (Li and Ligon)
• Do individuals share risk with others who have similar risk preferences or different risk preferences? Using risk-sharing network data in Ghana, this paper finds that an individual’s risk pooling community features more diverse risk preferences than their network neighborhood. (Putman)
• In Thailand, when one household experiences a significant health shock, it propagates to other linked households via the village network. Specifically, upstream businesses close to the under-insured households experience reduced local sales and increased inventories, and workers closer to the underinsured households in the labor network have reduced earnings. (Kinnan et al)

Session 4E: Politicians & Autocrats

• Why do autocrats implement technologies increasing transparency? A new video monitoring system of elections in Russia reduces voter turnout and results in fewer votes cast for the incumbent while increasing citizens’ trust in the electoral institutions. (Faikina)
• This paper analyzes the effects of a historical natural experiment in Java where, as remuneration, chiefs in one half of a region were awarded cultivation rights over village rice land but not in the other. Using field survey data, analysis reveals that higher land rents cause positive chief performance and economic development by attracting higher quality leaders and aligning incentives with that of villagers. (Lim)
• Using data on Indian state elections, an analysis showed that political parties respond to competition in elections prior to the election by selecting higher quality candidates and reallocating resources to swing constituencies. (Shaukat)

If you would like to suggest a correction to any of the above summaries, please email CEGA Events and Communications Associate, Dustin Marshall, at dmarsh1231@berkeley.edu.